Harley Spaceframe

History

The idea of a universal structural system in which a limited number of comparatively small components could be combined to produce a variety of forms has engaged the attention of architects and engineers since the late 1800s when Thomas Edison and Alexander Graham Bell first began their experiments in the USA.  Initially, they were concerned with nautical and aeronautical frameworks, but also developed an interest in structures for building towers and enclosing space.  

They soon discovered three-dimensional triangulated networks, but the task of mechanically connecting members with their axes coincident, cost-effectively, eluded them.  The inevitable congestion occurring when eight or more members met at a node created a problem they were unable to solve.

The challenge was taken up by American thinker and inventor Buckminster Fuller.  He began experiments to use lightweight, three-dimensional frameworks to enclose space.  He also ventured into automobile design.  His Dymaxion motorcar (1933) used a space frame structure much like early flying machines.  Interestingly his octet truss is still the most used system for highway-spanning signage in the United States.

The long-term success of space frames for building applications, however, depended on devising methods of connecting members in a three-dimensional array at each node.  The design of an economical node connector was the challenge.  Over time numerous systems were developed, some are depicted below.

The widespan building industry required solutions that were not only aesthetically pleasing, but also cost-competitive. 

The German engineer, Dr Max Megeringerhausen patented the widely-used MERO system in 1943.  This required a cast and machined steel spherical node with threaded entry points for strut connection.  The struts were fitted with captured bolts.

Another system used a similar node, but the connection of struts was made by inserting the bolt into a bent tube welded into the strut or using a port and flexible drive system to make the connection, in practice this proved problematic, as seen:

These systems employing cast and machined hubs and cast, machine and welded end pieces on struts were ideal in theory, but expensive in application.  There was also a very large weight penalty generated by the connecting componentry.  

Whilst these structures appealed aesthetically, MERO was the only system that established a strong market presence, and it is still being used. 

Stephan de Chateau produced three systems between 1957 and 1960 which used hollow spherical nodes with an entry port that enabled connecting bolts to be inserted from inside the sphere.

A hollow sphere used in this way has to be very heavy to resist both compressive and tension forces and is expensive to manufactures.  It wasn’t particularly successful and is no longer used.

All of these systems had some success, but the weight and manufacturing cost of the connecting assemblies, which also required welding operations to connect components to the tubular struts was a cost penalty resulting in relatively small turnovers. 

Architects liked the aesthetics of networked structures, but high cost limited their use in all but prestigious projects. 

There is, of course, the additional problem of all consulting engineers wishing to design their own structures, rather than specifying an “off-the-shelf” product (there are negative fee implications).

Other than the weight penalty associated with discrete node-based systems, the cutting of materials produced by continuous industrial processes into short lengths and then using heavy components and many hours of labour to connect them lacks logic and results in costs that most projects could not accommodate. 

This was the state of play when Australian Architect and industrial designer, Edwin Codd, began his experiments to develop three-dimensional structural systems that could compete with traditional widespan structures based on trusses and columns or portals.

His strategy involved:

• Eliminating node connectors and using the actual truss members to obtain joint integrity with simple site bolting.  This was the most inventive step since it completely eliminated the need for heavy and expensive node connectors.
• Using components such as roll-formed channels in long lengths, which could be fully processed on line and then delivered directly to the site.
• Manufacturing the shorter web members on line, including manufacturing the tube, end forming, cutting to length and colour coating ready for delivery.

In some markets, a continual process from flat steel coil to finished members ready for delivery and site assembly took place on one, uninterrupted production line.  

His inventions are supported by three international patents; Series 80 (1984), Series 85 (1986), and Series 90 (1987).

The isolated groups working on space frames soon identified a need for research centres to develop ideas and undertake testing.

The most notable of these was the University of Surrey’s Space Structures Research Centre.  They began to organise international conferences on space structures.  The centre was founded by Professor Z. S. Makowski in 1963.  It is part of the department of Civil and Structural Engineering at the university. 

Professor Hoshyar Nooshin replaced Professor Makowski in 1971 and remained in this position for 28 years.  The current director is Professor Gerard Parke.

Many years after this initial work began and products such as the Mero system were well accepted, Australian architect Edwin Codd investigated other possibilities.  

His focus was on reducing weight, manufacturing complexity and using continuous on line methods of production.  He reasoned that lipped channels and steel tubes could be produced on line and be delivered to the site for assembly without any further processing. 

Testing of these structures confirmed the validity of this approach; failures only occurred when members buckled.  Importantly, the upper and lower chords had a reduced buckling length of approximately 0.7 resulting in a more efficient use of material.

In theory, having the axes of members entering the connecting coincident was desirable but all practical methods of meeting this objective had two major downsides, namely: 

     – expensing componentry; and

     – massive weight increase.

It is apparent that for most of the systems illustrated, not only is the node componentry heavy and expensive to produce, requiring casting, machining and welding, but there is a lack of flexibility in the sizing of struts – the diameter is determined by the connecting components, which have to be standardised and therefore have capacity well beyond that required in lightly-loaded areas of the frame. 

It was against this background that Australian architect Edwin Codd developed the first of three systems for which he was granted patents.

     – Harley Series 80 (patent granted 1984) 

     – Harley Series 85 (patent granted 1986)

     – Harley Series 90 (patent granted 1987)

He reasoned that in all of the above systems that if the node connector was eliminated, and the members entering the joint were connected directly with simple bolting, that the resulting structure would be approximately half the weight of the other systems available.

In addition, labour costs would decrease dramatically because all processing could take place on line including cutting to length and end forming, hole punching and decorative surface treatments.

Fortunately, at the time his experiments began, the Australian steel industry led many developments internationally.  It was well-equipped in numerous locations in every state to produce high-strength, cold-rolled zinc-coated components in a range of gauges.

The fifth international conference on space structures is to be held in August 2020 at Gilford in England.

When Codd presented a paper at the third International Space Structures Conference, the centre at that time was headed by Professor Makowski.  Much concern was expressed by engineers present about the validity of the systems presented by Codd.

Three PhD students at the Imperial College of London finally developed the theory to support the assumptions which had formed the basis of the many structures already built.  Codd subsequently presented papers across North and South America, Europe, New Zealand, the Middle East and South-East Asia.

Licences were granted in all these markets.  However, the patents have long since expired.  Are these systems still being used?  Possibly not.  Architectural and Engineering Consulting practices earn their fees based on developing responses unique to each project.  Using proprietary systems for a major component of expenditure on a building project is often not considered an attractive business proposition.  No doubt, the better systems will survive and there is every chance that as we are in the race for space, structures using lightweight repetitive units may yet make a comeback.